Lipsync troubleshooting

Most AI lip sync issues in Workroom come from source file quality or model mismatch. Here is how to fix the most common ones.

Generate button stays gray

The Generate button only activates when both a source file and a voiceover are selected. If it stays gray, at least one input is empty.

Click Video (or Image if using LTX Lipsync or Lipsync Image Labs) and select a file. Then click Voiceover and select an audio track. See Create a lip-sync video for the step-by-step flow.

Lips don't sync accurately

Sync errors usually come from voiceover pacing. Long pauses, very fast speech, or background noise in the audio make lip tracking harder. Trim silence from the start and end of your voiceover before uploading. You can do this in Voice Studio or any audio editor.

Source video quality also matters — shaky footage or a face that moves frequently out of center reduces accuracy. A steady video with a face in the center works best.

Face is not detected

If generation produces no visible lip movement, the model may not have found a usable face in the source. Common causes:

Multiple faces in the frame
Face too small or partially out of frame
Low contrast or very dark footage

Use a video with one clearly visible face, close to the center, with good lighting. See Prepare source files for more input guidance.

Output quality is lower than expected

LTX Lipsync is the fastest model and works with photo input. For higher quality from a photo, switch to Lipsync Image Labs. For video input, use Lipsync Video Labs or Sync.co Lipsync.

See Lipsync models for a full comparison.

Processing takes longer than expected

Processing time scales with voiceover length. A 30-second voiceover generates faster than a 5-minute one. There's no way to speed it up within the app.